| graph-classification-on-cifar10-100k | GAT | #17 | Accuracy (%): 65.48 |
| graph-classification-on-dd | GAT | #46 | Accuracy: 73.109±3.413 |
| graph-classification-on-enzymes | GAT | #3 | Accuracy: 78.611±1.556 |
| graph-classification-on-imdb-b | GAT | #3 | Accuracy: 84.250±2.062 |
| graph-classification-on-nci1 | GAT | #12 | Accuracy: 85.109±1.107 |
| graph-classification-on-nci109 | GAT | #13 | Accuracy: 82.560±0.601 |
| graph-classification-on-proteins | GAT | #33 | Accuracy: 76.786±1.670 |
| graph-property-prediction-on-ogbg-code2 | GAT | #11 | Test F1 score: 0.1569 ± 0.0010Ext. data: No… |
| graph-regression-on-esr2 | GAT | #5 | R2: 0.666±0.000RMSE: 0.510±0.666 |
| graph-regression-on-f2 | GAT | #4 | R2: 0.886±0.000RMSE: 0.343±0.886 |
| graph-regression-on-kit | GAT | #4 | R2: 0.833±0.000RMSE: 0.443±0.833 |
| graph-regression-on-lipophilicity | GAT | #3 | RMSE: 0.536±0.020R2: 0.820±0.014 |
| graph-regression-on-parp1 | GAT | #4 | R2: 0.921±0.000RMSE: 0.353±0.921 |
| graph-regression-on-pgr | GAT | #6 | R2: 0.681±0.000RMSE: 0.546±0.681 |
| graph-regression-on-zinc-100k | GAT | #8 | MAE: 0.463 |
| graph-regression-on-zinc-full | GAT | #16 | Test MAE: 0.078±0.006 |
| heterogeneous-node-classification-on-acm | GAT | #11 | Micro-F1: 92.19Macro-F1: 92.26 |
| heterogeneous-node-classification-on-dblp-2 | GAT | #11 | Micro-F1: 93.39Macro-F1: 93.83 |
| heterogeneous-node-classification-on-freebase | GAT | #9 | Macro-F1: 40.74Accuracy: 65.26 |
| heterogeneous-node-classification-on-imdb | GAT | #11 | Micro-F1: 64.86Macro-F1: 58.94 |
| molecular-property-prediction-on-esol | GAT | #6 | RMSE: 0.540±0.027R2: 0.930±0.007 |
| molecular-property-prediction-on-freesolv | GAT | #5 | RMSE: 0.791±0.101R2: 0.959±0.011 |
| node-classification-on-brazil-air-traffic | GAT (Velickovic et al., 2018) | #7 | Accuracy: 0.382 |
| node-classification-on-chameleon-60-20-20 | GAT | #19 | 1:1 Accuracy: 63.9 ± 0.46 |
| node-classification-on-citeseer | GAT | #33 | Accuracy: 72.5 ± 0.7%Training Split: fixed 20 per nodeValidation: YES |
| node-classification-on-citeseer-05 | GAT | #13 | Accuracy: 38.2% |
| node-classification-on-citeseer-1 | GAT | #14 | Accuracy: 46.5% |
| node-classification-on-citeseer-60-20-20 | GAT | #31 | 1:1 Accuracy: 67.20 ± 0.46 |
| node-classification-on-citeseer-with-public | GAT | #26 | Accuracy: 72.5 ± 0.7% |
| node-classification-on-cora | GAT | #33 | Accuracy: 83.0% ± 0.7%Training Split: fixed 20 per node… |
| node-classification-on-cora-05 | GAT | #13 | Accuracy: 41.4% |
| node-classification-on-cora-1 | GAT | #14 | Accuracy: 48.6% |
| node-classification-on-cora-3 | GAT | #15 | Accuracy: 56.8% |
| node-classification-on-cora-60-20-20-random | GAT | #30 | 1:1 Accuracy: 76.70 ± 0.42 |
| node-classification-on-cora-with-public-split | GAT | #23 | Accuracy: 83.0 ± 0.7% |
| node-classification-on-cornell-60-20-20 | GAT | #26 | 1:1 Accuracy: 76.00 ± 1.01 |
| node-classification-on-europe-air-traffic | GAT (Velickovic et al., 2018) | #5 | Accuracy: 42.4 |
| node-classification-on-film-60-20-20-random | GAT | #27 | 1:1 Accuracy: 35.98 ± 0.23 |
| node-classification-on-flickr | GAT (Velickovic et al., 2018) | #8 | Accuracy: 0.359 |
| node-classification-on-genius | GAT | #25 | Accuracy: 55.80 ± 0.87 |
| node-classification-on-non-homophilic | GAT | #26 | 1:1 Accuracy: 76.00 ± 1.01 |
| node-classification-on-non-homophilic-1 | GAT | #26 | 1:1 Accuracy: 71.01 ± 4.66 |
| node-classification-on-non-homophilic-13 | GAT | #14 | 1:1 Accuracy: 81.53 ± 0.55 |
| node-classification-on-non-homophilic-2 | GAT | #29 | 1:1 Accuracy: 78.87 ± 0.86 |
| node-classification-on-non-homophilic-4 | GAT | #18 | 1:1 Accuracy: 63.9 ± 0.46 |
| node-classification-on-non-homophilic-6 | GAT | #22 | 1:1 Accuracy: 61.09±0.77 |
| node-classification-on-pattern-100k | GAT | #8 | Accuracy (%): 75.824 |
| node-classification-on-penn94 | GAT | #19 | Accuracy: 81.53 ± 0.55 |
| node-classification-on-ppi | GAT | #16 | F1: 97.3 |
| node-classification-on-pubmed | GAT | #43 | Accuracy: 79.0 ± 0.3%Training Split: fixed 20 per nodeValidation: YES… |
| node-classification-on-pubmed-003 | GAT | #12 | Accuracy: 50.9% |
| node-classification-on-pubmed-005 | GAT | #13 | Accuracy: 50.4% |
| node-classification-on-pubmed-01 | GAT | #13 | Accuracy: 59.6% |
| node-classification-on-pubmed-60-20-20-random | GAT | #36 | 1:1 Accuracy: 83.28 ± 0.12 |
| node-classification-on-pubmed-with-public | GAT | #22 | Accuracy: 79.0% |
| node-classification-on-squirrel-60-20-20 | GAT | #23 | 1:1 Accuracy: 42.72 ± 0.33 |
| node-classification-on-texas-60-20-20-random | GAT | #32 | 1:1 Accuracy: 78.87 ± 0.86 |
| node-classification-on-usa-air-traffic | GAT (Velickovic et al., 2018) | #4 | Accuracy: 58.5 |
| node-classification-on-wisconsin-60-20-20 | GAT | #29 | 1:1 Accuracy: 71.01 ± 4.66 |
| node-property-prediction-on-ogbn-arxiv | GAT+label reuse+self KD | #16 | Test Accuracy: 0.7416 ± 0.0008Ext. data: No… |
| node-property-prediction-on-ogbn-arxiv | GAT+label+reuse+topo loss | #21 | Test Accuracy: 0.7399 ± 0.0012Ext. data: No… |
| node-property-prediction-on-ogbn-products | GAT with NeighborSampling | #44 | Test Accuracy: 0.7945 ± 0.0059Ext. data: No… |
| node-property-prediction-on-ogbn-proteins | GAT + labels + node2vec | #7 | Ext. data: NoTest ROC-AUC: 0.8711 ± 0.0007… |